Okay, first off, I've been trying to make games for the past 5 years, if not longer, and it's plain to see, I'm doing something wrong! (that and not enough motivation)

Heres the rough steps I take when making a game, with reasons why I'm doing it; If someone has any suggestions, PLEASE do so.
1) Get a general Idea; Write a little bit about it (If I write too much about it, I abandon the game cause of the complexities)

2) Do some art which hopefully will make it thru to beta stage (Since I've Never had a good artist willing to help, I'm stuck making my own, so I make some "official" art so I know that I can make the rest easily enough)
3) Spend a few days looking at engines to pick the one I want (usually it's Irrlicht or Panda3D)
4) Start Programming

<--Then I stop due to lack of interest or I never make it this far because of art, theres a few times I get past this-->
5) Make some more assets and add them into the game

6) Program some more features

Repeat 5-6 until I get what I'm aiming for; I never get to 6 because the "more assets" are usually art that I find a hard time doing and no one else wants to do so it doesn't end up so well -.-'

Any hints of what I may be doing wrong? other than loosing interest or giving up too fast >.<

Have you accomplished anything ? Are you trying 2D or 3D ? I am assuming 3D by the choice of Panda3D.

You must have a generalized theme you are clearly interested in ? If so, Press on building your art or 3D models tool box until and idea comes to mind.

I built a 3D Space exploration type game with 1 Player Ship, 1 Enemy Ship, 1 Docking Station.

My game Started out with exploring a section of a planets surface looking for randomly placed Clues / Plans to build other space ships or mining equipment.

The mission was to assembly a set of plans. Since these plans had to be assembled from the clues, thise gave me enough time to design other ships and build on my game engine that I had designed.

So based upon your order of the game design process I Would re-order as such

1 - (3) Pick a language and SDK and stick with it.

2 - (2) Built your art work / 3D Models. ( Vehicles are easy in that the need Little or No Animation )

3 - (4) Start Programming. Every Program has a certain set of elements. Graphics Intialization, Disk operations, Player input ( Keyboard / Joystick/\ / Game pad. Build that and then save this a seperate utility program to add to or build on later.

4 - (1) Get an Idea. Fledge it out. Maybe work it as a table top game or something where you can see how it might work.

5 - (5) Make any additional assets

6 - (6) Do some more programming. Even if you give it up, you have something going.

Yeah, art is a problem for most of us developers with no sense of it and ADHD as well. I don't have much advice to give other than to try to use some debug crappy asset till you develop the features you want.

Thankfully, I've overcame that little problem. Art isn't a problem as it once was...sure HUD will be a pain in the rear, but I know enough to make the basic HUD and I can spend hours later trying to perfect it;

1 - (3) Pick a language and SDK and stick with it.

Basically it's Irrlicht or Panda3D, depending on what I'm aiming for; This newest project (lets hope it works out!) will be done in Panda3D; primary reason: it's the only engine I got to step 6 with >.< the only reason why that game was canceled was because I did a very stupid thing and told Panda to do collision detection using a ray-polygon with every polygon that wasn't the character...and that was a few years ago now...